I have a plenty of FASTQ files (FASTQ is a standard for storing the output of high-throughput sequencing instruments such as the Illumina Genome Analyzer) and need to convert them to GTF format (gtf - file format used to hold information about gene structure which significant feature is that it is validatable: given a sequence and a GTF file, one can check that the format is correct. This significantly reduces problems with the interchange of data between groups).

I assume that (if those formats are popular files' standards) there must exist some software that easily converts information from FASTQ files to GTF file. Did anyone hear about any such (is possible open-source) software? I am familiar with R statistical package and SAS and can learn Python on a fly.

Thanks for any help.

  • 2
    $\begingroup$ A quick Google search yields many hits, not to mention the feedback from below. Please edit your question and include more about what you have, what you don't have, what you're trying to do, what resources you have to work with, what your time constraints are, what your ultimate goals are, etc. There is absolutely no guarantee that anyone here will be able to answer it, or that it will be on-topic here, as "Too Broad" questions will get closed. $\endgroup$
    – MattDMo
    Sep 8, 2015 at 20:44

4 Answers 4


You should really read about both these file formats. As swbarnes mentioned, FASTQ and GTF hold different kind of information. GTF stores the annotation of a reference sequence. For example a GTF for a genome sequence will have the information about the locations of features such as genes, transcripts, exons, start codon etc.

FASTQ stores the sequence of a read obtained from sequencing along with the quality scores corresponding to each position.

As mentioned by others, asking for interconversion of these file formats, makes no sense.

I guess what you are asking for is "How to obtain novel annotations with a FASTQ file at hand?"

This also depends on what you want to annotate.

The sixth column in a GTF file refers to a score; you can assign expression values to different features. You can calculate expression using the read counts. If these are RNAseq reads then expression can be measured using packages like tophat-cufflinks, RNAstar or some others.

If you are doing ChIP-Seq then you can generate a GTF with a new feature called TFBS (transcription factor binding site) and annotate the locations. A popular package used for ChIP-Seq analysis is MACS, which takes your reads and outputs the TFBS in the form of a BED file which also stores co-ordinates. You can convert BED to GTF. You can also assign scores based on the read counts at different TFBS.

If you do not have a reference genome or if the annotation of the reference genome is incomplete, then you should first assemble your reads. If you have a reference genome then you can go for a reference guided assembly of the transcripts to obtain novel transcripts or splice variants; Cufflinks does this.

If you do not have a reference genome then you should go for de-novo assembly of your transcriptome and annotate the transcriptome for start codons or other features of processed transcripts. Velvet and Trinity are popular packages that do de-novo assembly.


Your question is not very clear. What should the contents of your GTF file be? Typically, GTF files contain information about where exons are located in a set of DNA sequences. Determining the location and exon/intron structure of genes is no simple technical task (i.e. a "conversion" as stated in your question), but rather a large area of active research. Annotating genes involves the use of statistical modeling (ab initio gene predictors), aligning experimental evidence (ESTs, cDNAs, and potentially Illumina RNA-Seq reads), and in some cases manual refinement of computational predictions. If you're working with a model organism like human or mouse or fruit fly, reliable GTF files are readily available for download from public databases. If you're not working with a model organism, then you many have quite a lot of work cut out to annotate a genome from scratch.

Or perhaps you want to annotate new alternatively spliced isoforms for known genes?

Without more information, it will be difficult for us to help you understand how to process your raw data (Illumina reads in FASTQ format) into a GTF file that addresses some biological question you're interested in.


A fastq contains sequences. A gtf contains coordinates of where features like exons fall in a reference sequence. You can't interconvert them, that makes no sense.

  • $\begingroup$ So it is impossible to create a gtf file when I have information only from fastq file? $\endgroup$ Sep 8, 2015 at 19:53
  • $\begingroup$ If this is an novel organism, and you have a very nice large set of RNAseq reads, you could in theory align it to your reference, and make a gtf from that. But in general, no, you want to download an appropriate gtf for your organism, not make your own. $\endgroup$
    – swbarnes2
    Sep 8, 2015 at 20:04
  • $\begingroup$ No, I want to create my own file. I do have a very nice large set of RNAseq reads and the gtf file does not exist. That's the question :) $\endgroup$ Sep 8, 2015 at 20:09
  • $\begingroup$ You can't just convert. You have to align to your reference for starters, as Daniel says, making a gtf from scratch is a big project, to do it right will be a lot of work. But if that's what you need to do, then that's what you need to do. $\endgroup$
    – swbarnes2
    Sep 8, 2015 at 21:29

The Tuxedo Suite (Tophat, Bowtie and cufflinks) used to to process RNA_seq data, assuming that is the origin of your .fastq files, should work for you.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.